Jim Nasby

NOTE! I stuck a comment in the documentation about `# >= 10 AND #

There's no need to do any work in the case that jsquery is NULL.

EXPLAIN BUFFERS should report skip-list statistics

1

Currently, EXPLAIN (BUFFERS,VERBOSE) does not output any information about skip list performance. I think that's something that anyone looking at buffer data would want to know.

Ability to redeploy only changed files

11

Use case 1: While developing patches, changes are made to deploy, revert or verify files, possibly including "old" ones. The rebase command should be able to determine what change to...

It would be nice to have an edit command that would essentially do the same thing that `sqitch add -e CHANGE` does (except just opening the editor).

Add rename command

1

Sometimes you decide your original plan for a name was a bad idea. It would be nice if you could rename a change. To be clear: I am NOT asking...

Change pgtap_version() from numeric?

2

Now that we're past 1.0.0, it'd be helpful to be able to do patch releases (ie, to fix issue #209). We can't do that with pgtap_version returning numeric, unless we...

Currently the `coltap` and `hastap` tests can't run in parallel, presumably due to a name conflict. It'd be nice to allow them to run in parallel.

Remove support for pre-9.3

10

I'd like to officially remove support for pre-9.3 versions of Postgres, due to issues in #128. 9.2 went [EOL 6 months ago](https://www.postgresql.org/support/versioning/). https://travis-ci.org/decibel/pgtap/builds/358206497 shows the failures that #128 is getting...

col_is_pk variants

2

There should also be `(name, name, name)` and `(name, name, name[])` versions of `col_is_pk`. Those versions already exist for `col_is_unique()`.